Reader Comments

Post a new comment on this article

Referee Comments: Referee 1 (Denis Nash)

Posted by PLOS_ONE_Group on 26 Jun 2007 at 12:55 GMT

Reviewer 1's Review (Denis Nash)

“This is a well-written paper describing an important development in surveillance for HIV clinical outcomes and behaviors. The authors provide a thorough justification for the need to develop and implement a new surveillance system that is more representative of people accessing HIV care than ASD, PSD, SHAS, and SHDC. I recommend that the paper be published after suitable revision.

Specific comments:
• The overall goal of the paper should be stated in the abstract and the body of the text. Is it to justify and describe the design of the MMP?

• The paper refers to the MMP alternately as a new surveillance system that needs to be put in place and one that already is in place. It should be made clear to the reader up front that this is a newly implemented existing surveillance system.

• What are the overarching goals of the MMP? These are referred to often throughout the manuscript, but I felt it would be important, after justifying the need for a new system, to explicitly state/enumerate the goals and objectives (primary and secondary) of the new surveillance system.

• What is the target population? There also should be a very specific definition of the target population. Often the text states that it is "persons in care", which I think is not specific and detailed enough. There should be some component of time included. Often the text refers to persons not in care as one of the focus populations for the MMP. Given earlier definitions, this can be confusing/contradictory. Some statement like: "All persons who accessed care at any time HIV care and treatment sites in the US in a given year". Also, the time period of data abstraction is never described. Is it for a given calendar year? Or all history that's in the chart? What about time frame for interviews?

• Population-based vs. representative. These terms are used throughout the manuscript as desirable attributes that the MMP is aiming to achieve. However, I'm not sure that "population-based" applies here if the term population is meant to mean all persons in the US population. This is a select population (depending on the target population definition) of persons accessing care and treatment for HIV). It seems to me that the most important factor here is that the MMP sample be 'representative' of it's target population. Not population-based. The use of this term should be clarified/justified or dropped in favor of another, more accurate term (e.g., representative of target population).

Tables and figures
• It would be good to have a figure/graphic showing the sampling stages and also describe the frame for each stage. The reader should come away with a very clear understanding of the sampling strategy and how estimates will be weighted to target populations, etc.

• Maps of past sites and current sites would be helpful to demonstrate the geographic representation that is referred to. Right now, the only way to get a sense of this is from the names of the states/sites in the tables. For geographically challenged readers (e.g., New Yorkers), a map would be much better.

• Information in Table 1 could also probably be better conveyed with a map that indicates the number of patients and clinics. The names could be included in a legend. Also, since the "facility-based" term applies to all of the cohorts, it could be stated in the title. You could also omit non-US cohorts and make the map US-based, as that is the focus of the paper. Should the authors decide to use maps instead of tables to present the geographic information, it would be desirable to also include, for comparison, a map of the MMP sites, overlaid with case surveillance data of PLWHA. This would help substantiate the authors' assertion that there is a need for a more geographically representative surveillance system AND that the MMP achieves this.

• Table 2 could also be restructured to convey the message more succinctly and save space. The first row should include the race/ethnicity distribution from national case surveillance (perhaps with a range for the lowest and highest reported proportion for the reporting jurisdictions; the second row should be the race/ethnicity distribution of all US HIV cohorts combined. Instead of listing each individual cohort, a range could be included for the lowest and highest proportion from the cohorts for a given race/ethnicity. However, ASD, SHAS, and SHDC should be listed out separately. Alternatively, a graphical representation of the distributions could be presented. The title of table 3 should list the time period and population (e.g., 33 states) for case reporting data. As it stands, while all the relevant information is in the table, it's not presented in a way that is easily for the reader to take in.

• Information from table 3 could be included on the MMP map mentioned above instead of in it's present form. Also, while it's nice to know the sample size, it would also be helpful to see the expected precision for each state/site and for the country as a whole.

• Table 4 does not really include data elements, but rather domains of data elements. I wonder if it's useful to be more specific, or include a link to a table that is more comprehensive in the appendix. I think it would also be important to mention the target population and time frame for each MMP component.

• Table 5. It is not clear over what time frame these outcomes are assessed. Is it for the first year of care? Is it at least one laboratory outcome for the entire time patients were in care? I think this table could be omitted and the info stated in the text.”

N.B. These are the general comments made by the reviewer when reviewing this paper in light of which the manuscript was revised. Specific points addressed during revision of the paper are not shown.