Inclusion, characteristics and methodological limitations of systematic reviews in doctoral theses: A cross-sectional study of all universities in Sweden

Abstract

Intro

A systematic review (SR) attempts to find, assess and summarize all the empirical evidence to answer a specific research question. We aim to explore to what extent reviews are included in doctoral theses from all universities with a medical faculty in Sweden, and to describe the type, topic and assess the methodological quality of the reviews.

Methods

Duplicate assessors independently searched local and national repositories for doctoral theses published in 2021 within all seven medical faculties in Sweden, and categorized identified reviews based on review type, topic, and methodological quality using AMSTAR-2.

Results

5.4% (45/852) of all doctoral theses included a review, and 1.3% (45/3461) of all included studies were reviews. Of these, two thirds (31) were SRs and the rest (14) were broader ‘big picture’ reviews. The most common topics were interventions (42%) and exposure/etiology (32%), with no reviews of diagnostic tests. The majority of the SRs had very low (71%) or low (19%) quality, and few reached a high (7%) or moderate (3%) quality. The most common issues were limitations with protocols, limited search strategies, and failure to account for risk of bias in drawn conclusions.

Conclusions

Few doctoral students included SRs in their theses, and the few SRs included in doctoral theses generally had a low quality. There is no consensus on the appropriate proportion of doctoral thesis including a SR. We argue that conducting a SR within a doctoral thesis can reduce redundant, harmful and unethical research, identify knowledge gaps, and help the doctoral student obtain important skills to conduct and use research.