The paper itself has lengthy discussion on the flaws of the research it's examining.

The studies reviewed examined the possible impact of SPPA on users or their partners using cross-sectional designs, with one study also using longitudinal research methods. Of course, retrospective cross-sectional designs cannot be used to draw causal conclusions 51 about any associations between SPPA and potential outcomes because they are measured simultaneously; it might be difficult to ascertain whether individuals perceived their pornography use to be problematic before or after they experi- enced negative outcomes. Moreover, the longitudinal study used a two-wave design and a much smaller subset (n ¼ 106) of the original sample (N ¼ 1,215), which substantially limits causality- related analyses, so findings are likely to be tentative at best.

Nine of the 10 studies reported evidence that SPPA had a detrimental impact on individuals or their partners. However, some important methodologic issues must be considered. First, SPPA and its impact were often assessed using a single-item measurement, which research suggests is an adequate measure- ment of complex constructs. 4,52 If an individual’s experience is multidimensional (ie, physiologic, behavioral, and cognitive), then it might be challenging for the individual to convey this using a single item, and assumptions can be made that omit potentially important information. Second, some studies used under-defined concepts and definitions; for example, Levin et al 19 used a single-item measurement to assess impaired functioning resulting from SPPA but did not provide a definition of functioning, so it is uncertain whether the researchers were measuring the same construct for all participants

Third, three studies18,20,21 suggested that individuals’ values and morals associated with their pornography use might have contributed to their perceived pornography addiction, and Prause et al20 further suggested that conflict with their held values might have led to their distress. Therefore, SPPA might actually result from a conflict in values rather than pornography use per se.

Research that examined the impact of SPPA on the partners of self-perceived pornography addicts found that they experienced several negative effects such as feelings of betrayal, shame, and isolation. These effects were attributed to the behavior of the self- perceived pornography addict. However, research investigating the effects of pornography use has shown that women who attribute their partners’ pornography use to an inadequacy about themselves experience a greater level of distress.53 None of the studies reviewed considered the characteristics of the partners of self-perceived pornography addicts, yet negative outcomes can be affected by factors such as thinking styles and attitudes (eg, how we perceive information), which can lead to these feelings of inadequacy.

There also were concerns regarding the measurements used to make conclusions about the impact of SPPA. Many relied on adapted and non-validated measurements that were not neces- sarily theoretically driven and were derived from a non-clinical sample; thus, the findings are difficult to generalize. For example, Twohig et al21 used a median cutoff (58%) from a non- clinical sample to determine an arbitrary level of problematic cognitive and behavioral outcomes of SPPA.

I could keep going, but I think that's enough for this post - read the "Correlates and Possible Outcomes of SPPA" and "Limitations" sections of the paper you linked.