{"id":196,"date":"2014-07-18T08:38:49","date_gmt":"2014-07-18T08:38:49","guid":{"rendered":"https:\/\/staffblogs.le.ac.uk\/bayeswithstata\/?p=196"},"modified":"2025-02-26T13:21:38","modified_gmt":"2025-02-26T13:21:38","slug":"network-meta-analysis-in-stata","status":"publish","type":"post","link":"https:\/\/staffblogs.le.ac.uk\/bayeswithstata\/2014\/07\/18\/network-meta-analysis-in-stata\/","title":{"rendered":"Network Meta-analysis in Stata"},"content":{"rendered":"

Regrettably clinical trials are often conducted that are too small to provide definitive evidence of the benefit of one treatment over another. In order to compensate for this, there has been a lot of work over the last twenty years on meta-analysing, or combining, the results from different trials. There is an example of a Bayesian meta-analysis in chapter 6 of the \u2018Bayesian analysis with Stata<\/em>\u2019.<\/p>\n

A complication that\u00a0frequently arises in the meta-analyses of clinical trials is that different trials make slightly different treatments comparisons. So, in one trial a drug might be compared with a placebo and in another trial the same drug might be compared with an active control. Sorting out such a mixture of comparisons is often referred to as network meta-analysis.<\/p>\n

In a recent talk that I attended, the process of network meta-analysis was likened to measuring the height difference between three people. Let\u2019s name the three people A, B and Placebo. In one experiment we compare A with Placebo and find that A in 2cm taller. Then in a separate study we compare B with Placebo and find that B is 3cm taller. Although we have never compared A and B directly we can deduce that B is taller than A by 1cm. The idea is simple but is it valid?<\/p>\n

\"heights\"<\/a><\/p>\n

Suppose that subsequently we learn that the heights of A and Placebo were compared when they were both seated, while B and Placebo were compared when they were both standing. We can still calculate that the difference is 1cm but what does it represent? In fact\u00a0the deduced difference\u00a0has no useful meaning. You might assume that at least we know that B is the taller than A, but if it were the case that seated differences are always half those measured when standing, then A is actually taller than B.<\/p>\n

One has to be very careful with analogies but we can easily see the consequences for treatments compared in clinical trials. Only when the trials comparing drug A with Placebo are conducted under exactly the same conditions as the trials comparing drug B with Placebo, can we be certain of obtaining a meaningful measure of the difference between A and B. Here, the ‘same conditions’ refers to anything that alters the measurement of treatment difference and might include factors such as, the inclusion\/exclusion criteria, the\u00a0exact form\u00a0of the placebo treatment\u00a0or the trial methodology; rarely will we be confident that these conditions are met.<\/p>\n

Does the requirement for identical trial conditions matter? This is a difficult question to answer, certainly different conditions are likely to introduce bias into the indirect comparison of A and B but how big that bias will be will vary from problem to problem. There is however one sense in which this potential bias is very important. Organizations such as NICE (National Institute for Health and Care Excellence) in the UK have started to use network meta-analysis when considering which treatments to finance under the NHS (National Health Service) and a bias in such a comparison could have very important consequences.<\/p>\n

For a much more detailed discussion of the assumptions needed for valid network meta-analysis see the recent articles by Donegan et al (Research Synthesis Methods 2013;4:291-323) and Jansen and Naci (BMC Medicine 2013;11:159).<\/p>\n

There is a report that was produced for NICE that is a good starting point for anyone considering a Bayesian network meta-analysis as it contains a detailed review, followed by several examples together with the WinBUGS code required for\u00a0their Bayesian analysis. ( http:\/\/www.nicedsu.org.uk\/TSD2%20General%20meta%20analysis%20corrected%2015April2014.pdf<\/a> \u00a0).<\/p>\n

One of the examples considered in the NICE report concerns a meta-analysis of the incidence of diabetes amongst people in trials of different treatments for hypertension. Concern is that the use of diuretics to treat high blood pressure might increase the risk of developing diabetes.<\/p>\n

The report structures the data as,<\/p>\n

+--------------------------------------------------------------------+\r\n\r\n| time\u00a0\u00a0 t_1 \u00a0\u00a0r_1\u00a0\u00a0\u00a0 n_1\u00a0\u00a0 t_2\u00a0\u00a0 r_2\u00a0\u00a0\u00a0 n_2\u00a0\u00a0 t_3\u00a0\u00a0 r_3\u00a0\u00a0\u00a0 n_3\u00a0\u00a0 na |\r\n\r\n|--------------------------------------------------------------------|\r\n\r\n|\u00a0 5.8\u00a0\u00a0\u00a0\u00a0 1\u00a0\u00a0\u00a0 43\u00a0\u00a0 1081\u00a0\u00a0\u00a0\u00a0 2\u00a0\u00a0\u00a0 34\u00a0\u00a0 2213\u00a0\u00a0\u00a0\u00a0 3\u00a0\u00a0\u00a0 37\u00a0\u00a0 1102\u00a0\u00a0\u00a0 3 |\r\n\r\n|\u00a0 4.7\u00a0\u00a0\u00a0\u00a0 1\u00a0\u00a0\u00a0 29\u00a0\u00a0\u00a0 416\u00a0\u00a0\u00a0\u00a0 2\u00a0\u00a0\u00a0 20\u00a0\u00a0\u00a0 424\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0 2 |\r\n\r\n|\u00a0\u00a0\u00a0 3\u00a0\u00a0\u00a0\u00a0 1\u00a0\u00a0 140\u00a0\u00a0 1631\u00a0\u00a0\u00a0\u00a0 2\u00a0\u00a0 118\u00a0\u00a0 1578\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0 2 |\r\n\r\n|\u00a0 3.8\u00a0\u00a0\u00a0\u00a0 1\u00a0\u00a0\u00a0 75\u00a0\u00a0 3272\u00a0\u00a0\u00a0\u00a0 3\u00a0\u00a0\u00a0 86\u00a0\u00a0 3297\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0 2 |\r\n\r\n|\u00a0\u00a0\u00a0 4\u00a0\u00a0\u00a0\u00a0 1\u00a0\u00a0 302\u00a0\u00a0 6766\u00a0\u00a0\u00a0\u00a0 4\u00a0\u00a0 154\u00a0\u00a0 3954\u00a0\u00a0\u00a0\u00a0 5\u00a0\u00a0 119\u00a0\u00a0 4096\u00a0\u00a0\u00a0 3 |\r\n\r\n|--------------------------------------------------------------------|\r\n\r\n|\u00a0\u00a0\u00a0 3\u00a0\u00a0\u00a0\u00a0 1\u00a0\u00a0 176\u00a0\u00a0 2511\u00a0\u00a0\u00a0\u00a0 4\u00a0\u00a0 136\u00a0\u00a0 2508\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0 2 |\r\n\r\n|\u00a0 4.1\u00a0\u00a0\u00a0\u00a0 1\u00a0\u00a0 200\u00a0\u00a0 2826\u00a0\u00a0\u00a0\u00a0 5\u00a0\u00a0 138\u00a0\u00a0 2800\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0 2 |\r\n\r\n|\u00a0\u00a0\u00a0 1\u00a0\u00a0\u00a0\u00a0 1\u00a0\u00a0\u00a0\u00a0 8\u00a0\u00a0\u00a0 196\u00a0\u00a0\u00a0\u00a0 6\u00a0\u00a0\u00a0\u00a0 1\u00a0\u00a0\u00a0 196\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0 2 |\r\n\r\n|\u00a0 3.3\u00a0\u00a0\u00a0\u00a0 2\u00a0\u00a0 154\u00a0\u00a0 4870\u00a0\u00a0\u00a0\u00a0 4\u00a0\u00a0 177\u00a0\u00a0 4841\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0 2 |\r\n\r\n|\u00a0\u00a0\u00a0 3\u00a0\u00a0\u00a0\u00a0 2\u00a0\u00a0 489\u00a0\u00a0 2646\u00a0\u00a0\u00a0\u00a0 5\u00a0\u00a0 449\u00a0\u00a0 2623\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0 2 |\r\n\r\n|--------------------------------------------------------------------|\r\n\r\n|\u00a0 4.5\u00a0\u00a0\u00a0\u00a0 2\u00a0\u00a0 155\u00a0\u00a0 2883\u00a0\u00a0\u00a0\u00a0 5\u00a0\u00a0 102\u00a0\u00a0 2837\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0 2 |\r\n\r\n|\u00a0 4.8\u00a0\u00a0\u00a0\u00a0 2\u00a0\u00a0 399\u00a0\u00a0 3472\u00a0\u00a0\u00a0\u00a0 5\u00a0\u00a0 335\u00a0\u00a0 3432\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0 2 |\r\n\r\n|\u00a0 3.1\u00a0\u00a0\u00a0\u00a0 2\u00a0\u00a0 202\u00a0\u00a0 2721\u00a0\u00a0\u00a0\u00a0 6\u00a0\u00a0 163\u00a0\u00a0 2715\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0 2 |\r\n\r\n|\u00a0 3.7\u00a0\u00a0\u00a0\u00a0 2\u00a0\u00a0 115\u00a0\u00a0 2175\u00a0\u00a0\u00a0\u00a0 6\u00a0\u00a0\u00a0 93\u00a0\u00a0 2167\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0 2 |\r\n\r\n|\u00a0 3.8\u00a0\u00a0\u00a0\u00a0 3\u00a0\u00a0 \u00a070\u00a0\u00a0\u00a0 405\u00a0\u00a0\u00a0\u00a0 4\u00a0\u00a0\u00a0 32\u00a0\u00a0\u00a0 202\u00a0\u00a0\u00a0\u00a0 5\u00a0\u00a0\u00a0 45\u00a0\u00a0\u00a0 410\u00a0\u00a0\u00a0 3 |\r\n\r\n|--------------------------------------------------------------------|\r\n\r\n|\u00a0\u00a0\u00a0 4\u00a0\u00a0\u00a0\u00a0 3\u00a0\u00a0\u00a0 97\u00a0\u00a0 1960\u00a0\u00a0\u00a0\u00a0 4\u00a0\u00a0\u00a0 95\u00a0\u00a0 1965\u00a0\u00a0\u00a0\u00a0 5\u00a0\u00a0\u00a0 93\u00a0\u00a0 1970\u00a0\u00a0\u00a0 3 |\r\n\r\n|\u00a0 5.5\u00a0\u00a0\u00a0\u00a0 3\u00a0\u00a0 799\u00a0\u00a0 7040\u00a0\u00a0\u00a0\u00a0 4\u00a0\u00a0 567\u00a0\u00a0 7072\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0 .\u00a0 \u00a0\u00a0\u00a0\u00a0.\u00a0\u00a0\u00a0 2 |\r\n\r\n|\u00a0 4.5\u00a0\u00a0\u00a0\u00a0 3\u00a0\u00a0 251\u00a0\u00a0 5059\u00a0\u00a0\u00a0\u00a0 4\u00a0\u00a0 216\u00a0\u00a0 5095\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0 2 |\r\n\r\n|\u00a0\u00a0\u00a0 4\u00a0\u00a0\u00a0\u00a0 3\u00a0\u00a0 665\u00a0\u00a0 8078\u00a0\u00a0\u00a0\u00a0 4\u00a0\u00a0 569\u00a0\u00a0 8098\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0 2 |\r\n\r\n|\u00a0 6.1\u00a0\u00a0\u00a0\u00a0 3\u00a0\u00a0 380\u00a0\u00a0 5230\u00a0\u00a0\u00a0\u00a0 5\u00a0\u00a0 337\u00a0\u00a0 5183\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0 2 |\r\n\r\n|--------------------------------------------------------------------|\r\n\r\n|\u00a0 4.8\u00a0\u00a0\u00a0\u00a0 3\u00a0\u00a0 320\u00a0\u00a0 3979\u00a0\u00a0\u00a0\u00a0 6\u00a0\u00a0 242\u00a0\u00a0 4020\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0 2 |\r\n\r\n|\u00a0 4.2\u00a0\u00a0\u00a0\u00a0 4\u00a0\u00a0 845\u00a0\u00a0 5074\u00a0\u00a0\u00a0\u00a0 6\u00a0\u00a0 690\u00a0\u00a0 5087\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0\u00a0\u00a0 .\u00a0\u00a0\u00a0 2 |\r\n\r\n+--------------------------------------------------------------------+<\/strong><\/span><\/pre>\n
So the first study compared three treatments (na=3) and followed its subjects for time=5.8 years. In that study, the treatments (t_1, t_2, t_3) were 1 (diuretics), 2 (placebo) and 3 (beta-blockers); 1081 (n_1) patients were treated with diuretics and of these 43 (r_1) were diagnosed with diabetes during the follow up period.<\/p>\n
Bias might occur if, for example, diuretics tend to be compared with placebo in trials that recruit more elderly patients or perhaps more over-weight patients and if treatment differences are larger in such patients.\u00a0Such unmeasured differences are a particular problem when they act multiplicatively.<\/p>\n
A simple model for these data assumes a proportional hazards structure for the incidence of diabetes. This would imply that the probability, p, of developing diabetes in a follow up of length t years would be given by,
\nlog(-log(1-p)) = X\u03b2 + log H(t)
\nwhere H is the cumulative hazard and X is the design matrix. This complementary log-log transformation is commonly used for such data, although we have no way of confirming the proportional hazards assumption that underpins it.<\/p>\n
Before embarking on the Bayesian analysis it is helpful to conduct a simple exploration of the data and to analyse them using maximum likelihood. This will give us a good indication of the answer that a Bayesian analysis would produce under non-informative priors.<\/p>\n
A simple complementary log-log model is,
\nlog(-log(1-pij)) = \u03b1i + \u03b4j + log Hi(ti)
\nwhere i denotes study and j denotes treatment. Since we have no information about the form of Hi we might as well combine that term with the study effects \u03b1i and write the model as
\nlog(-log(1-pij)) = \u03b1i + \u03b4j<\/p>\n
In the report the authors chose the model
\nlog(-log(1-pij)) = \u03b1i + \u03b4j \u2013 \u03b41i + log(ti)
\nwhere \u03b41i is the first treatment applied in study i. This formulation is over-parameterized and makes it appear that we have information about the cumulative hazard that we cannot actually justify. The inclusion of \u03b41i changes the interpretation of the study effect but has no impact on the treatment differences. As the ordering of treatments within\u00a0a trial is arbitrary it is hard to see why the authors chose to include \u03b41i. Probably this over-parameterization explains why the authors had to run their Bayesian analysis for so long before achieving convergence.<\/p>\n
Let us consider a maximum likelihood analysis of the model,
\nrij ~ B(pij,nij)
\nlog(-log(1-pij)) = \u03bc + \u03b1i + \u03b4j
\nUnder the constraint that \u03b11=\u03b41=0
\nThis is a standard generalized linear model that we can fit using the Stata command glm.<\/p>\n
To fit the model we need to stack the data. So starting with the layout shown above,
\ngen id = _n
\nstack id time t_1 r_1 n_1 id time t_2 r_2 n_2 id time t_3 r_3 n_3 \/\/\/
\n, into( id time t r n) clear
\ndrop if t == .
\nglm r i.id i.t, fam(bin n) link(cloglog)<\/p>\n
The estimates for the differences from treatment 1 (diuretics) are,<\/p>\n
------------------------------------------------------------------------------<\/span><\/strong><\/pre>\n
\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 |\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 OIM<\/span><\/strong><\/pre>\n
\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 r |\u00a0\u00a0\u00a0\u00a0\u00a0 Coef.\u00a0\u00a0 Std. Err.\u00a0\u00a0\u00a0\u00a0\u00a0 z\u00a0\u00a0\u00a0 P>|z|\u00a0\u00a0\u00a0\u00a0 [95% Conf. Interval]<\/span><\/strong><\/pre>\n
-------------+----------------------------------------------------------------<\/span><\/strong><\/pre>\n
\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 2\u00a0 |\u00a0 -.2469671\u00a0\u00a0 .0562828\u00a0\u00a0\u00a0 -4.39 \u00a0\u00a00.000\u00a0\u00a0\u00a0 -.3572794\u00a0\u00a0 -.1366548<\/span><\/strong><\/pre>\n
\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 3\u00a0 |\u00a0\u00a0 -.056982\u00a0\u00a0 .0557689\u00a0\u00a0\u00a0 -1.02\u00a0\u00a0 0.307\u00a0\u00a0\u00a0 -.1662869\u00a0\u00a0\u00a0\u00a0 .052323<\/span><\/strong><\/pre>\n
\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 4\u00a0 |\u00a0 -.2530259\u00a0\u00a0 .0536676\u00a0\u00a0\u00a0 -4.71\u00a0\u00a0 0.000\u00a0\u00a0\u00a0 -.3582125\u00a0\u00a0 -.1478392<\/span><\/strong><\/pre>\n
\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 5\u00a0 |\u00a0\u00a0 -.358473\u00a0\u00a0 .0533383\u00a0\u00a0\u00a0 -6.72\u00a0\u00a0 0.000\u00a0\u00a0\u00a0 -.4630142\u00a0\u00a0 -.2539318<\/span><\/strong><\/pre>\n
\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 6\u00a0 |\u00a0 -.4527086\u00a0\u00a0 .0630479\u00a0\u00a0\u00a0 -7.18\u00a0\u00a0 0.000\u00a0\u00a0\u00a0 -.5762803\u00a0\u00a0 -.3291369<\/span><\/strong><\/pre>\n
------------------------------------------------------------------------------<\/span><\/strong><\/pre>\n
Under proportional hazards, the coefficients represent log hazard ratios and so treatment 1 (diuretics) does have the highest rate of diabetes with all but treatment 3 (beta-blockers) showing a significantly lower rate.<\/p>\n
With any binomial analysis we need to be concerned about over-dispersion and in this case the Pearson X2\/df ratio is 2.4 suggesting possible mild over-dispersion but nothing extreme. The deviance residual plot confirms this impression,
\npredict rd , deviance
\npredict xb , xb
\nscatter rd xb, yline(0)<\/p>\n
\"NMresid\"<\/a><\/p>\n
The residuals show no pattern and most are within the range \u00b12 as we would hope. All in all, this simple model seems quite adequate to describe the data. We might neatly summarize the results in a plot of the treatment coefficients.<\/p>\n
parmest , norestore
\ngen id = _n
\ntwoway (scatter estimate id in 24\/28) \/\/\/
\n(rcap min95 max95 id in 24\/28) , leg(off) \/\/\/
\nxlabel(24 “Placebo” 25 “b-blocker” 26 “CCB” 27 “ACE” 28 “ARB”) \/\/\/
\nyline(0) text(0.05 27 “Diuretic”) xtitle(Treatment)<\/p>\n
\"NMCI\"<\/a><\/p>\n
Anyone thinking of undertaking a Network Meta-analysis is Stata should visit the website
\nhttp:\/\/www.mtm.uoi.gr\/index.php\/stata-routines-for-network-meta-analysis<\/a>
\nwhere they will find a collection of Stata programs for graphically representing networked trials. The recent Stata Journal article \u2018Indirect treatment comparison\u2019 by Miladinovic et al 2014:14(1)76-86 should also be of interest.<\/p>\n
In my next posting I will return to the\u00a0diabetes data\u00a0and describe how we can combine Stata with WinBUGS to produce a Bayesian network meta-analysis of these data similar to that suggested in the NICE report.<\/p>\n","protected":false},"excerpt":{"rendered":"
Regrettably clinical trials are often conducted that are too small to provide definitive evidence of the benefit of one treatment over another. In order to compensate for this, there has been a lot of work over the last twenty years on meta-analysing, or combining, the results from different trials. There is an example of a […]<\/p>\n","protected":false},"author":134,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[3,24,4],"class_list":["post-196","post","type-post","status-publish","format-standard","hentry","category-uncategorized","tag-bayesian","tag-network-meta-analysis","tag-stata"],"_links":{"self":[{"href":"https:\/\/staffblogs.le.ac.uk\/bayeswithstata\/wp-json\/wp\/v2\/posts\/196","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/staffblogs.le.ac.uk\/bayeswithstata\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/staffblogs.le.ac.uk\/bayeswithstata\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/staffblogs.le.ac.uk\/bayeswithstata\/wp-json\/wp\/v2\/users\/134"}],"replies":[{"embeddable":true,"href":"https:\/\/staffblogs.le.ac.uk\/bayeswithstata\/wp-json\/wp\/v2\/comments?post=196"}],"version-history":[{"count":5,"href":"https:\/\/staffblogs.le.ac.uk\/bayeswithstata\/wp-json\/wp\/v2\/posts\/196\/revisions"}],"predecessor-version":[{"id":205,"href":"https:\/\/staffblogs.le.ac.uk\/bayeswithstata\/wp-json\/wp\/v2\/posts\/196\/revisions\/205"}],"wp:attachment":[{"href":"https:\/\/staffblogs.le.ac.uk\/bayeswithstata\/wp-json\/wp\/v2\/media?parent=196"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/staffblogs.le.ac.uk\/bayeswithstata\/wp-json\/wp\/v2\/categories?post=196"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/staffblogs.le.ac.uk\/bayeswithstata\/wp-json\/wp\/v2\/tags?post=196"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}