Evaluation and the Health Professions

Although it is increasingly common, metaanalysis is still a relatively new addition to the toolbox of research strategies. Partially because of its newness and partially because its early enthusiasts may have overstated its presumed contributions, a number of common misunderstandings have developed about the meta-analytic approach. These misunderstandings have to do with the way that metaanalysis is conducted and interpreted and its relation to primary research and narrative review. This article aims to highlight these misunderstandings and offer an alternative view of meta-analysis. It is hoped that metaanalytic research of the future will be a partnership with primary research and that metaanalysis will be valued not so much for its definitiveness as for the creative speculations it allows about patterns that appear across studies.

Summarizing the results of research related to identifying potential predictors of hospital readmissions has been difficult because of conflicting results across studies. Using the techniques of meta-analysis, the results from 44 studies were examined in the present study. OveralI the mean readmission rate was 27%, with significant differences based on patient diagnosis. Although diagnosis, age, initial length of hospital stay, and prior use of hospital resources were related to readmission, the strength of the relationship is trivial. Combining data from 12 intervention studies designed to reduce readmission indicated that the overall treatment effect was not significant. Further research is needed to determine demographic, clinical, and social predictors of readmission if strategies are to be developed to reduce readmission and the resulting health care costs.

Medical record review (MRR) is often used in clinical research and evaluation, yet there is limited literature regarding best practices in conducting a MRR, and there are few studies reporting interrater reliability (IRR) from MRR data. The aim of this research was twofold: (a) to develop a MRR abstraction tool and standardize the MRR process and (b) to examine the IRR from MRR data. This study introduces the MRR-Conduction Model, which was used to implement a MRR, and examines the IRR between two abstractors who collected preinjury medical and psychiatric, incident-related medical and postinjury head symptom information from the medical records of 47 neurologically injured workers. Results showed that the percentage agreement was ≥85% and the unweighted κ statistic was ≥.60 for most variables, indicating substantial IRR. An effective and reliable MRR to abstract medical-related information requires planning and time. The MRR-Conduction Model is proposed to guide the process of creating a MRR.

This article offers some guidelines for interpreting and evaluating meta-analytic reviews of research. The fundamental goals of metaanalysis are to combine results across studies to yield an overall estimate of effect and to compare effects between studies in order to understand moderating factors. Suggestions are made for what readers should look for in a meta-analysis, and a discussion is provided of several issues that are not often explicitly addressed: choice of unit of analysis, fixed versus random effects, the meaning of heterogeneity, determination of when contrasts are appropriate, and the choice of measure of central tendency. We recommend that readers adopt a skeptical attitude about the results of meta-analysis, particularly when only complex analyses are reported.

The FACIT (Functional Assessment of Chronic Illness Therapy) translation methodology has been in use for nearly 10 years and, during the course of that time, has emphasized a universal translation approach that includes multicountry review, the use of qualitative and quantitative methods in testing, and the exploration of new methods such as differential item functioning (DIF) analysis using item response theory to evaluate item equivalence. The FACIT translation methodology aims to establish equivalence of meaning and measurement between different country versions through the use of the decentered model of translation and advanced statistical methods.

Choosing a specialty within the occupation of medicine is problematic given the number of available specialty choices. A need exists for the development and modification of inventories to be used to help students with selecting a medical specialty. Furthermore, ratings of job satisfaction, regardless of specialty, are mixed with some physicians regretting choosing medicine as a career. Despite its use in medical specialty counseling, research is lacking regarding the Medical Specialty Preference Inventory (MSPI). We examined the predictive validity of the MSPI and compared job satisfaction of physicians (N= 51) whose specialty was correctly (i.e., congruent) versus incorrectly (i.e., incongruent) predicted by the MSPI. The MSPI correctly predicted the specialty for 33%. Congruent physicians had higher job satisfaction than incongruent physicians. Results of this study may help individuals who provide medical specialty counseling or who are involved in designing tools and implementing programs to assist students with medical specialty decision making.

An increasing number of research studies over the last three decades suggest that a wide range of substance and process addictions may serve similar functions. The current article considers 11 such potential addictions (tobacco, alcohol, illicit drugs, eating, gambling, Internet, love, sex, exercise, work, and shopping), their prevalence, and co-occurrence, based on a systematic review of the literature. Data from 83 studies (each study n = at least 500 subjects) were presented and supplemented with small-scale data. Depending on which assumptions are made, overall 12-month prevalence of an addiction among U.S. adults varies from 15% to 61%. The authors assert that it is most plausible that 47% of the U.S. adult population suffers from maladaptive signs of an addictive disorder over a 12-month period and that it may be useful to think of addictions as due to problems of lifestyle as well as to person-level factors.