Screening Tools for Neurocognitive Impairment in HIV/AIDS
Screening Tools for Neurocognitive Impairment in HIV/AIDS
With the prevalence of milder forms of HAND increasing, and limited resources available for formal neuropsychological examinations, there is a critical need to be able to screen and identity people with HAND. Our current systematic review identified 51 peer-reviewed articles that focused on screening for neurocognitive impairments associated with HIV/AIDS. Of these, about two-thirds (31/51) used a criterion or reference to characterize the performance of 21 unique screening instruments. About half of the 31 studies examined the performance of the HDS or the IHDS as screening tests. Fifty-five percent of studies focused on detecting neurocognitive impairments, 35% on HAD and 15% in identifying MND or ANI. Only 16% of studies employed the latest 2007 HAND criteria.
The results of this systematic review suggest that the HDS and IHDS are not ideal tools for identifying a range of neurocognitive impairment. From the meta-analysis on the HDS and IHDS, we were able to determine that these tools have generally poor (0.48) and moderate (0.62) pooled sensitivities, respectively, in identifying a range of neurocognitive impairment. We also identified a potential relationship between the reference test and criterion with the range of sensitivities on the HDS and the IHDS, that is those studies using the most comprehensive neuropsychological test battery as the 'gold standard' criterion had the lowest (and similar) sensitivity values, whereas those that were considered 'short' or 'medium' in time (and comprehensiveness) resulted in much larger ranges of sensitivities. The association between lower sensitivity (ability to identify true positives) and large reference battery was significant for the HDS (P < 0.05). These findings suggest that a longer, more comprehensive neuropsychological battery may be a more stringent and appropriate reference test. Future studies evaluating sensitivity may want to consider using a large neuropsychological battery as a reference test to classify participants in a more discerning way. Studies using the HDS to identify a range of neurocognitive impairment reported higher sensitivities when higher cut-offs were used or demographically adjusted T-scores were used.
The review identified 10 other screening tools with adequate sensitivities (≥0.75). Of these, four tools or combinations of tests were used to detect HAND conditions and overall neurocognitive impairment (as opposed to impairment in specific domains) and they used a 'gold-standard' neuropsychological battery as the reference test or criterion. These four tools include the CogState, the Screening Algorithm, the paired Hopkins Verbal Learning Test and WAIS-III Digit Symbol combination and the paired Hopkins Verbal Learning Test and Grooved Pegboard Non-Dominant Hand combination. It is important to note that the other studies defined neurocognitive impairment in different ways and evaluated these conditions using different criteria; thus, the comparison of efficacy across all studies needs to be carried out with caution given these limitations. In addition to these four tests, Becker et al. reported a slightly lower sensitivity (0.72) for the Computer Assessment of Mild Cognitive Impairment (CAMCI) against a comprehensive neuropsychological battery. These are the five tools identified by our review that show some promise to be used to identify HAND in a clinical setting, although they warrant further study and validation.
In the 20 studies that met review criteria and are reported in Appendix B, http://links.lww.com/QAD/A370, one other potentially useful tool, The Montreal Cognitive Assessment (MoCA), was highlighted. This tool is worth note because of its rising popularity in clinical settings, although the reliability and validity to detect HAND remains to be determined.
The review also highlighted key methodological issues in this body of literature. First, in order to properly determine whether a screening tool can detect the target condition, it is imperative that the reference test properly classifies the condition. Only 19 studies of the 31 (61%) used the 'gold-standard' neuropsychological battery as the reference test. Furthermore, only seven (23%) studies used more comprehensive reference tests (>90 min) and functional status was measured objectively in only nine studies (29%). In addition, only five studies used the most recent 2007 criteria to classify HAND. The QUADAS assessment reported that only 17 of the 31 (54%) were using representative samples, noting a severe limitation in the applicability of the studies. For cognitive screening tools to be properly evaluated, and for these evaluations to inform clinical practice, more stringent methodological procedures are needed.
The results of this systematic review provide insights into the performance of the current screening instruments available for HAND and identify key methodological issues that need to be addressed in future studies. In this current context of mild cognitive impairment associated with HIV, it is clear that improved screening tools could go a long way in improving the care, quality of life and study of treatment interventions for individuals living with HIV and AIDS.
Discussion
With the prevalence of milder forms of HAND increasing, and limited resources available for formal neuropsychological examinations, there is a critical need to be able to screen and identity people with HAND. Our current systematic review identified 51 peer-reviewed articles that focused on screening for neurocognitive impairments associated with HIV/AIDS. Of these, about two-thirds (31/51) used a criterion or reference to characterize the performance of 21 unique screening instruments. About half of the 31 studies examined the performance of the HDS or the IHDS as screening tests. Fifty-five percent of studies focused on detecting neurocognitive impairments, 35% on HAD and 15% in identifying MND or ANI. Only 16% of studies employed the latest 2007 HAND criteria.
The results of this systematic review suggest that the HDS and IHDS are not ideal tools for identifying a range of neurocognitive impairment. From the meta-analysis on the HDS and IHDS, we were able to determine that these tools have generally poor (0.48) and moderate (0.62) pooled sensitivities, respectively, in identifying a range of neurocognitive impairment. We also identified a potential relationship between the reference test and criterion with the range of sensitivities on the HDS and the IHDS, that is those studies using the most comprehensive neuropsychological test battery as the 'gold standard' criterion had the lowest (and similar) sensitivity values, whereas those that were considered 'short' or 'medium' in time (and comprehensiveness) resulted in much larger ranges of sensitivities. The association between lower sensitivity (ability to identify true positives) and large reference battery was significant for the HDS (P < 0.05). These findings suggest that a longer, more comprehensive neuropsychological battery may be a more stringent and appropriate reference test. Future studies evaluating sensitivity may want to consider using a large neuropsychological battery as a reference test to classify participants in a more discerning way. Studies using the HDS to identify a range of neurocognitive impairment reported higher sensitivities when higher cut-offs were used or demographically adjusted T-scores were used.
The review identified 10 other screening tools with adequate sensitivities (≥0.75). Of these, four tools or combinations of tests were used to detect HAND conditions and overall neurocognitive impairment (as opposed to impairment in specific domains) and they used a 'gold-standard' neuropsychological battery as the reference test or criterion. These four tools include the CogState, the Screening Algorithm, the paired Hopkins Verbal Learning Test and WAIS-III Digit Symbol combination and the paired Hopkins Verbal Learning Test and Grooved Pegboard Non-Dominant Hand combination. It is important to note that the other studies defined neurocognitive impairment in different ways and evaluated these conditions using different criteria; thus, the comparison of efficacy across all studies needs to be carried out with caution given these limitations. In addition to these four tests, Becker et al. reported a slightly lower sensitivity (0.72) for the Computer Assessment of Mild Cognitive Impairment (CAMCI) against a comprehensive neuropsychological battery. These are the five tools identified by our review that show some promise to be used to identify HAND in a clinical setting, although they warrant further study and validation.
In the 20 studies that met review criteria and are reported in Appendix B, http://links.lww.com/QAD/A370, one other potentially useful tool, The Montreal Cognitive Assessment (MoCA), was highlighted. This tool is worth note because of its rising popularity in clinical settings, although the reliability and validity to detect HAND remains to be determined.
The review also highlighted key methodological issues in this body of literature. First, in order to properly determine whether a screening tool can detect the target condition, it is imperative that the reference test properly classifies the condition. Only 19 studies of the 31 (61%) used the 'gold-standard' neuropsychological battery as the reference test. Furthermore, only seven (23%) studies used more comprehensive reference tests (>90 min) and functional status was measured objectively in only nine studies (29%). In addition, only five studies used the most recent 2007 criteria to classify HAND. The QUADAS assessment reported that only 17 of the 31 (54%) were using representative samples, noting a severe limitation in the applicability of the studies. For cognitive screening tools to be properly evaluated, and for these evaluations to inform clinical practice, more stringent methodological procedures are needed.
The results of this systematic review provide insights into the performance of the current screening instruments available for HAND and identify key methodological issues that need to be addressed in future studies. In this current context of mild cognitive impairment associated with HIV, it is clear that improved screening tools could go a long way in improving the care, quality of life and study of treatment interventions for individuals living with HIV and AIDS.
Source...