Hello Everyone,
I'm new to MI and was hoping to get some advice on how to impute "plausible" start and stop dates for HIV Antiretroviral Drugs (ideally using IVEware). I'm currently running up against at least two issues. First, I'm having trouble forcing start dates to come before stop dates. Obviously, a patient can't stop using a drug before they started taking it. So this is a highly undesirable state of affairs. IVEware has a BOUNDS statement that should accomplish this task but it doesn't appear to be working properly. Or perhaps my syntax is wrong. I submitted to IVEware the following: "BOUNDS START_DATE(<STOP_DATE);". Interestingly, this statement allows me to impute start dates that come before stop dates when the start is missing and the stop is not missing or when both the start and stop are missing. Unfortunately though, it frequently imputes stops that come after starts when the start date is present and the stop date is missing. So IVEware seems to understand that the start date should be less than the stop date but doesn't seem to recognize that this also implies that the stop date should be greater than the start date. I tried altering the BOUNDS statement to read "BOUNDS STOP_DATE(>START_DATE)" but this just created a new problem that is essentially a mirror image of the old one. And it's not possible to specify both statements simultaneously. Is there any way to get IVEware to force all the start dates for drugs to come before the stop dates? Or is there any other software that can do this? Another problem is that even when starts come before stops, my imputed values seem implausible. This is the case because missing values are only partially missing. For example, suppose that I know a patient stopped taking a drug in May of 1998 but I don't know the particular day. Even if I give IVEware the available date information it will still impute values that are implausible. Thus, IVEware might guess that the patient stopped taking the drug on 3/13/1999. But this doesn't seem like a plausible guess, given what we know about the month and year the patient actually stopped taking the drug. (Maybe this doesn't really matter though. Perhaps it's O.K. that my values are implausible as long as they preserve an underlying pattern of relationships.) I had thought about imputing missing values for START_MONTH, START_DAY, START_YEAR, STOP_MONTH, STOP_DAY, STOP_YEAR and then using them to calculate the START_DATE and STOP_DATE. That would make the values for START_DATE and STOP_DATE more plausible. But it wouldn't necessarily constrain the START_DATE to come before the STOP_DATE. And also it seems unlikely that any covariate I might introduce into the data would be particularly predictive of, say, the day of the month or day when a patient stopped taking a drug. So I guess I'm at a bit of a loss. If anyone has any advice, I'd greatly appreciate it. Thanks, Paul Paul J. Miller, Ph.D. Research Scientist and Statistician Ontario HIV Treatment Network 1300 Yonge St., Suite 308 Toronto, Ontario M4T 1X3 Phone: (416) 642-6486 ext 232 Fax: (416) 640-4245 -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.utsouthwestern.edu/pipermail/impute/attachments/20060829/ee810dd3/attachment.htm From rarose <@t> email.unc.edu Thu Aug 31 10:48:22 2006 From: rarose <@t> email.unc.edu (Roderick A. Rose) Date: Thu Aug 31 10:49:07 2006 Subject: [Impute] Imputing "Plausible" Start and Stop Dates for HIV Antiretroviral Drugs In-Reply-To: <[email protected]> References: <[email protected]> Message-ID: <fc4afc854dd2bc0b61287...@rodmus2> Paul, My recommended solution is made under the (perhaps incorrect) assumption that what you are mainly interested in is the interval between the start and stop dates and not the actual stop and start dates themselves. Let the start date equal zero in every case (so it doesn't have to be imputed) and the interval is a count of days (or another unit) between zero and the stop date. You impute this interval. I've not used IVEware, so I'm not sure this will completely eliminate the problem (e.g., you might end up with negative intervals if the bounds statement really doesn't work well). Regarding the second issue of plausibility, I am curious if it is necessary to have precision in days; if you know it happened in May 1998, you can err on the side of the least undesirable bias (by making it either May 31 or May 1). This is an alternative to ignoring the known value and letting it impute a completely new and possibly unrelated value. (Or do both and see what happens, as many of us probably do). Best, Rod --On Tuesday, August 29, 2006 11:12 AM -0400 Paul Miller <[email protected]> wrote: > > > Hello Everyone, > > > > I'm new to MI and was hoping to get some advice on how to impute > "plausible" start and stop dates for HIV Antiretroviral Drugs (ideally > using IVEware). I'm currently running up against at least two issues. > First, I'm having trouble forcing start dates to come before stop dates. > Obviously, a patient can't stop using a drug before they started taking > it. So this is a highly undesirable state of affairs. > > > > IVEware has a BOUNDS statement that should accomplish this task but it > doesn't appear to be working properly. Or perhaps my syntax is wrong. I > submitted to IVEware the following: "BOUNDS START_DATE(<STOP_DATE);". > Interestingly, this statement allows me to impute start dates that come > before stop dates when the start is missing and the stop is not missing > or when both the start and stop are missing. Unfortunately though, it > frequently imputes stops that come after starts when the start date is > present and the stop date is missing. So IVEware seems to understand that > the start date should be less than the stop date but doesn't seem to > recognize that this also implies that the stop date should be greater > than the start date. I tried altering the BOUNDS statement to read > "BOUNDS STOP_DATE(>START_DATE)" but this just created a new problem that > is essentially a mirror image of the old one. And it's not possible to > specify both statements simultaneously. Is there any way to get IVEware > to force all the start dates for drugs to come before the stop dates? Or > is there any other software that can do this? > > > > Another problem is that even when starts come before stops, my imputed > values seem implausible. This is the case because missing values are only > partially missing. For example, suppose that I know a patient stopped > taking a drug in May of 1998 but I don't know the particular day. Even if > I give IVEware the available date information it will still impute values > that are implausible. Thus, IVEware might guess that the patient stopped > taking the drug on 3/13/1999. But this doesn't seem like a plausible > guess, given what we know about the month and year the patient actually > stopped taking the drug. (Maybe this doesn't really matter though. > Perhaps it's O.K. that my values are implausible as long as they preserve > an underlying pattern of relationships.) > > > > I had thought about imputing missing values for START_MONTH, START_DAY, > START_YEAR, STOP_MONTH, STOP_DAY, STOP_YEAR and then using them to > calculate the START_DATE and STOP_DATE. That would make the values for > START_DATE and STOP_DATE more plausible. But it wouldn't necessarily > constrain the START_DATE to come before the STOP_DATE. And also it seems > unlikely that any covariate I might introduce into the data would be > particularly predictive of, say, the day of the month or day when a > patient stopped taking a drug. > > > > So I guess I'm at a bit of a loss. If anyone has any advice, I'd greatly > appreciate it. > > > > Thanks, > > > > Paul > > Paul J. Miller, Ph.D. > Research Scientist and Statistician > Ontario HIV Treatment Network > 1300 Yonge St., Suite 308 > Toronto, Ontario M4T 1X3 > Phone: (416) 642-6486 ext 232 > Fax: (416) 640-4245 > > -- Roderick A. Rose Research Associate, DDTI, Biculturalism and Making Choices Evaluation Director, School Success Profile School of Social Work, UNC Chapel Hill 919-962-8826 // Fax 962-7557
