Table 2 Analysis of bias for each paper selected for review

Li et al. 2016 [8]

Not likely. All patients were selected based upon objective measurements of Hepatitis B infection or Liver fluke infection

Not likely, objective measurements of Hepatitis B infection or Liver fluke infection are standard and not subjective (i.e., there are clear markers of infection)

N/A

Generalizability could be an issue because patients with these factors were excluded from analysis: “those co-infected with HIV, hepatitis A, C, D and E, those with type I and type II diabetes,

those co-infected with Schistosoma japonicum, or Schistosoma mansoni or other parasites,

and those with alcoholic liver, autoimmune diseases, cholestasis, serious heart diseases and pregnant women.”

Shin et al. 1996 [22]

Not likely. All patients were selected based upon objective measurements of Hepatitis B infection. Assessment of liver fluke infection was not described, but selection based upon observation in stool samples is unlikely related to known status of Hepatitis infection

Liver fluke infection could have been misclassified. Details were not described in Methods section

N/A

Relative risk estimates are very high, but the confidence intervals are wide. A larger sample size could have strengthened the certainty of relationships

Shi et al. 2017 [21]

Not likely. All patients were selected based upon objective measurements of Hepatitis B infection or Liver fluke infection

Although the study used objective measurements of Hepatitis B infection or Liver fluke infection, there was not a true control group. Of the 20 carcinoma patients, only one patient was not co-infected with liver fluke

N/A

Small sample size and lack of a control group limits certainty of evidence